详解一个Python库，用于构建精美数据可视化web app，练习做个垃圾分类app

import streamlit as st

# markdown
st.markdown('Streamlit is **_really_ cool**.')

# 设置网页标题
st.title('This is a title')

# 展示一级标题
st.header('This is a header')

# 展示二级标题
st.subheader('This is a subheader')

# 展示代码，有高亮效果
code = '''def hello():
  print("Hello, Streamlit!")'''
st.code(code, language='python')

# 纯文本
st.text('This is some text.')

# LaTeX 公式
st.latex(r'''
  a + ar + a r^2 + a r^3 + \cdots + a r^{n-1} =
  \sum_{k=0}^{n-1} ar^k =
  a \left(\frac{1-r^{n}}{1-r}\right)
''')

streamlit run garbage_classifier.py

  You can now view your Streamlit app in your browser.

  Local URL: http://localhost:8501
  Network URL: http://192.168.10.141:8501

import streamlit as st
import pandas as pd
import numpy as np
df = pd.DataFrame(
        np.random.randn(50, 5),
        columns=('col %d' % i for i in range(5)))

# 交互式表格
st.dataframe(df)
# 静态表格
st.table(df)

st.metric(label="Temperature", value="70 °F", delta="1.2 °F")

st.json({
    'foo': 'bar',
    'stuff': [
        'stuff 1',
        'stuff 2',
    ],
})

chart_data = pd.DataFrame(
    np.random.randn(20, 3),
    columns=['a', 'b', 'c'])

st.line_chart(chart_data)

import matplotlib.pyplot as plt

arr = np.random.normal(1, 1, size=100)
fig, ax = plt.subplots()
ax.hist(arr, bins=20)

st.pyplot(fig)

option = st.selectbox(
    '下拉框',
    ('选项一', '选项二', '选项三'))

st.write('选择了：', option)

DATE_COLUMN = 'date/time'
DATA_URL = ('https://s3-us-west-2.amazonaws.com/'
            'streamlit-demo-data/uber-raw-data-sep14.csv.gz')

@st.cache
def load_data(nrows):
    data = pd.read_csv(DATA_URL, nrows=nrows)
    lowercase = lambda x: str(x).lower()
    data.rename(lowercase, axis='columns', inplace=True)
    data[DATE_COLUMN] = pd.to_datetime(data[DATE_COLUMN])
    return data

import base64

import requests
import streamlit as st
import pandas as pd
import numpy as np

add_selectbox = st.sidebar.selectbox(
    "图片来源",
    ("本地上传", "URL")
)

uploaded_file = None
img_url = None

if add_selectbox == '本地上传':
    uploaded_file = st.sidebar.file_uploader(label='上传图片')
else:
    img_url = st.sidebar.text_input('图片url')

cls_mode = {'严格模式': 0, '模糊模式': 1}
mode_name = st.sidebar.radio('分类模式', cls_mode)
mode = cls_mode[mode_name]

# 请求结果
img_base64 = None
if uploaded_file:
    st.image(uploaded_file, caption='本地图片')

    base64_data = base64.b64encode(uploaded_file.getvalue())
    img_base64 = base64_data.decode()

if img_url:
    st.image(img_url, caption='网络图片')

if img_base64 or img_url:
    cls_res = get_img_cls_res(img_base64, img_url, mode)
    lajitype_to_name = {0: '可回收物', 1: '有害垃圾', 2: '厨余垃圾', 3: '其他垃圾', 4: '无法识别'}

    if cls_res.status_code == 200:
        cls_df = pd.DataFrame(cls_res.json()['newslist'])
        cls_df['分类'] = cls_df.index.astype(str) + '-' + cls_df['name'] + '-' + cls_df['lajitype'].apply(lambda x: lajitype_to_name[x])
        cls_df['置信度'] = cls_df['trust']
        cls_df.set_index(["分类"], inplace=True)
        print(cls_df)
        st.bar_chart(cls_df[['置信度']])
    else:
        st.write(cls_res)

def get_img_cls_res(img_base64, img_url, mode):
    url = 'https://api.tianapi.com/txapi/imglajifenlei/index'
    headers = {
        'Content-Type': 'application/x-www-form-urlencoded'
    }
    body = {
        'key': 'APPKEY',
        'mode': mode
    }

    if img_base64:
        body["img"] = img_base64
    if img_url:
        body['imgurl'] = img_url

    response = requests.post(url, headers=headers, data=body)

    return response